ABSTRACT : Autonomous sailboats are a very interesting type of mobile robot, since their main source of energy is wind, a free, abundant and environmentally friendly resource. Therefore, they can show great potential in different applications, such as long-term navigation and marine monitoring. In many applications, sailboats can not touch land for long periods, so the energy efficiency of their different systems is essential. Spiking neural networks (SNNs) are an artificial neural network model that is being investigated in the robotics field to implement controllers. These networks have, in theory, several advantages that make them quite attractive for robotic controllers design, such as low response latency and low energy consumption in specialized hardware. These qualities make SNNs a very interesting strategy to implement navigation controllers for autonomous sailboats. In this thesis, we present the development of an autonomous navigation system (ANS) based on spiking neural networks (SNNs) to solve a sailboat control task. The main objective was to find out how is the performance of such a controller for a sailboat, when following straight trajectories but also during tacking and gybing maneuvers. To run the SNNs, we developed a simulation environment based on the BindsNET library and the USVSim simulator. For each SNN, we used the modulated spike time-dependent plasticity reinforcement learning rule (M-STDP), a two-layer feed-forward network topology, and a leaky integrate-and-fire neuron model. We also defined the control strategy, training scenario and testing scenario. We performed two main experiments: a simulated design space exploration to tune SNN-based ANS hyperparameters and a test of an SNN-based ANS on a real sailboat. The obtained results show that the implemented control strategy works to develop an effective SNN-based ANS for the chosen control task, which performs both in simulation and in a real environment.